9 Jul
2006
9 Jul
'06
3:02 p.m.
On 7/9/2006 at 9:20 PM ard at p850ug1.demon.co.uk wrote:
As I mentioned in my reply to Allison, it's worse
than that.
Oh, I understand that there are problems. And in those situations where
problems exist with the 8255 and your application, it's probably best to
use a different chip or implementation.
In a way, this reminds me of Dijkstra railing against the horribleness of
the IBM 1620. Yes, it's true that there is no way for a program to
determine the contents of a memory location and it's impossible to output
certain characters, and so forth. The 1620 had a terribly incomplete
instruction set. But a lot of 1620's were sold and a lot of productive
work got done on them nonetheless. Would I choose the 1620 architecture to
implement a new computer? No way. But warts and all, the 1620 was
useful--as is the 8255.
Cheers,
Chuck